dahart

entirely possible that it’s a placebo

Placebo Effect is not Dunning Kruger. Making incorrect assumptions or having incorrect beliefs is not Dunning Kruger either.

Possible is not certain. So if it’s not a placebo effect? If the author found a solution that is in fact a better root cause fix for themselves than sleeping pills? What does that make their comment? What does it make yours?

This is so typical of the software dev community.

IMO this is a worse assumption with less evidence than the one you’re complaining about, and it’s pure unsupported assumption and far more likely to be wrong. Distrust of doctors and “western medicine” is common enough to hear it all the time, I know more non-devs than devs who carry such beliefs. Unless you have statistics on devs compared to the general population, your personal sampling is entirely anecdotal, and biased.

At least the judges read a lot. It could be worse…

For emphasis? That’s your takeaway?

Correlation does not imply causation

This is not relevant to what I said. We are closer to ending sexism and racism due to legislation, which is causal. We took specific actions to try to end sexism and racism and we know they had a positive effect.

Artificially balancing preferential treatment by applying it badly in the wrong direction is a duct tape solution

I agree! It was meant to be a duct tape solution, from the start. It was meant to be temporary and end preferential treatment and go away.

it may have been justified after the Civil Rights Act

I agree with this too, the justification was probably much stronger 60 or 70 years ago. It is weaker now… and that might be an indicator of its success.

Applying it badly has happened, and it’s lame and sucks, but that’s not an inherent feature or outcome of the core idea. Some might rightly say it was predictable.

The big open question is whether turning off all affirmative action will close that remaining gap, or leave us with a little bit of persistent systemic bias, or worse, a backslide. If that does, the next question is what to do.

What about them? Being white doesn’t make people immune to drugs, nor guarantee anything, right? So what’s your point? Race advantage is something you can’t always see by looking at individuals, especially cherry picked examples. Bias is a statistical effect, so you may need to look at everyone in order to see it, unless we are talking about clear cases of oppression, inability to vote or slavery or violence or things like that. Most of that has been resolved, but subtle bias is still around.

Holy Dunning-Kruger effect…

Do you know they’re wrong? Please don’t invoke Dunning Kruger like this, it’s cliché and also wrong to do. There’s no indicator for whatever it is Dunning & Kruger showed, you cannot know if it applies to a single person. Their main plot showed a positive correlation between confidence and competence. Their paper has problems, their methodology has been rightly questioned, and some attempts to reproduce have failed. Plus keep in mind that, ironically, for people who are intimately familiar with the debate over DK, using it to essentially name-call someone backfires and has the opposite of the intended effect, it makes the name caller look confidently ignorant.

Saying you have a preference is creepy. Having one and making decisions based on that is absolutely normal and always has been, and is not immoral, contrary to your claim. The vast majority of marriages globally are same race marriages.

Okay, that’s true, it’s one of the available choices. That’s a rather extreme position to take, given the history of slavery, racism, and sexism in the U.S. We know for a fact that outcome discrepancies were caused by racism before 1965. The discrepancies have been on the decline since then. So, when exactly, did equal opportunity start? We didn’t have it in 1966 yet, despite the law. Did we have it in 1980? Outcomes have closed since then, so no, obviously not. When, then? What reason is there to believe outcomes discrepancies aren’t a problem, given that society generally believes they are?

I’d be very happy to discuss the nuances and relative merits. I agree that it’s time to have race blind admissions and stop preferential treatment. I’m arguing against the hyperbolic notion that affirmative action is inherently immoral. @rayiner is trying to cut off the idea that it had any upsides and was ever reasonable. I’m still trying to convey that racism != any and all types of discrimination.

You’re the one supporting a system of affirmative action that treats people unequally

No, and you’re intentionally escalating things and I don’t appreciate it. I’ve been more than clear already that I don’t necessarily support affirmative action, and that I’m aware it has downsides and have acknowledged multiple times that it can have unintended consequences.

Unequal treatment already existed before affirmative action. That is the reason that affirmative action exists, and it is what affirmative action intended to fix. You have yet to acknowledge that fact in this discussion.

Wrong, but not illegal.

Why is racial preference in partner wrong?? Who agrees with this anywhere?

I reject your presumptive framing that affirmative action is a “punishment”. If you help someone, you are punishing the entire world?

that’s the very thing we decided is immoral

Racism is the thing we decided is immoral, actions based on belief that one group is inferior to another, not any and all notions of groups for any reason.

Objectively, Black Americans are extremely privileged by virtue of being Americans. A redlined community in 1950s America was still better than my dad’s village in Bangladesh.

Whoa. They should be happy with what they have, and put up with racism and inequality here because there are poorer people somewhere else in the world? Some Blacks in 1950 couldn’t vote - they were “privileged”?

Thank you. Yes the whole problem and confounding factor with this argument is that racism exists, and emerged on its own. If we choose Laissez Faire, we already know the outcome, and it’s bad. The notion that we should do nothing to fix it and let the chips fall where they may would amount to institutionalized racism as opposed to institutionalized (positive) discrimination. Plus affirmative action can and should end, whereas racism does not without changes.

Exactly. It depends. It’s not a given.

The death penalty is an example of two wrongs making a right in the minds of many people. It’s also at the same time an example of two wrongs not making a right. Imprisoning people is wrong, unless the government does it? All so-called lawful punishments of individuals are a form of hope that two wrongs do make a right.

Whether affirmative action is a wrong is your presumption, and it’s hotly contested, absolutely not universally agreed upon, which makes your use of ‘two wrongs don’t make a right’ a straw man.

There are plenty of things wrong with preferential affirmative action, but I reject the framing that it’s evil, this is essentially an ad hominem attempting to quash any debate of the actual relative merits or the outcomes.

I feel like we can’t make forward progress if you refuse to acknowledge the reasons that history has happened the way it did. See Chesterton’s Fence.

Discriminating against an individual based on their race is a moral wrong

People pick partners based on race. Is that morally wrong?

Our system of morality and justice operates on individuals, not groups.

What do you mean? Class actions don’t exist? Did I imagine the 19th amendment?

we have defined treating an individual differently based on their race as a moral wrong. The act is wrong, regardless of your motivations

No, I don’t think so, You’re conflating racism with discrimination. We humans discriminate positively and negatively all the time. People discriminate based on race and gender and age for the purposes of picking a life mate or sexual partner, and society has exactly zero problem with that. Discriminating based on race is absolutely not a morally wrong act regardless of motivations. Intent matters, and the situation matters.

This talking point about all discrimination being wrong is fairly popular, and there are a lot of people who can’t see the difference between positive discrimination and negative discrimination. This is one reason I think affirmative action should end; it’s too subtle of a distinction and we were too successful at planting the idea that discrimination is bad.

You’re also conflating law with mores in your example. There are cases where stealing something is moral, regardless of whether it’s legal. There are also cases where stealing something is legal too, for example you can legally take back something stolen from you if you do it non-violently.

There are no “groups”

I was referring to humanity, to all of us.

Yes! Exactly. Affirmative action should naturally end itself if it works. The goal is to equalize opportunity and stop preferential treatment, not to hang on to preferential treatment.

Well said, it does need to stop eventually. Now might be the right time, even if we’re not all the way there, given all the problems and unintended consequences and all the backlash.

Yep. Maybe a little bit like a sports handicap, you’re not trying to give someone a better chance of winning than everyone else, you’re trying to give them equal chances based on historical data.

Gemini gave me an okay analogy: “If a pendulum is knocked permanently off-center by a structural bend or a magnetic field, the only way to center it again is to apply an equal and opposite force. Affirmative action policies attempt to apply this counter-weight to straighten the broader societal scale.”

I’m not sure but is there a question about whether we actually had pre-existing unfairness? Blacks got the legal right to vote in 1965. Before that was effectively institutionalized racism, fully entrenched cultural bias, right?

Yes this is true, that can and has happened. I don’t think it’s true as rule, but it’s a valid point. Still, the most important question is what is the alternative?

We are closer to ending sexism and racism than we were 70 years ago before any affirmative action existed, right? Who’s to say that all the backfiring and unintended consequences that have happened aren’t better than the parallel universe in which maybe slavery still exists? Would it have happened faster with no quotas or boosts for groups that were historically discriminated against? Maybe yes, that’s possible, but we can’t know. What we do know is that we’ve made forward progress, even if imperfectly.

The thing that tends to be forgotten when people talk about affirmative action being preferential treatment is that the previous ‘in’ group was already getting preferential treatment, and maybe still is. We did not start from a level playing field. The core idea is to try to balance the pre-existing preferential treatment out. Unless you can resolve the cause of that, an easier way to attempt to handle it is to try to artificially make the preferences equal. The idea is not to give preferential treatment as an outcome, it’s to remove preferential treatment. This is related to how discrimination is not racism, and why positive discrimination is different from negative discrimination.

It feels like a platitudinal aphorism to me, it’s trite and cliche and overused and was delivered as if it’s insightful in hopes of ending the debate. Google appears to agree vigorously that “two wrongs don’t make a right” is a platitude.

One problem with that platitude is that there are plenty of ways that two negative things can balance each other out or become positive in some way.

Aren’t you making deep cultural presumptions and imposing your own opinions? Affirmative action is controversial, and it’s had unintended consequences when used, but what’s the rationale for claiming it’s a “wrong” or “injustice”? Not everyone believes that.

You’re also deliberately ignoring the point that there might be no such thing as innocent third parties, but only people who aren’t aware they’re part of the problem, even if it’s subtle and unintended.

History already tried the Laissez Faire approach, and it didn’t work. Chips fell in a bad place. This isn’t about proving wrongdoing or punishment, it’s about acknowledgement of a problem, and reflection and self-improvement as a group.

If only it were actually that simple and fixed with platitudes. Culture and belief is passed on from generation to generation, otherwise we wouldn’t have had the persistent problem in the first place. How do we fix a problem that we already have historical evidence that when left alone doesn’t go away on its own, that people who weren’t the original cause of the problem still, in fact, perpetuated it? How can we know people have nothing to do with the problem today, given that there are still discrepancies in outcome? I’m not defending Affirmative Action, the question at hand is what’s the alternative proposal?

This is institutionalized racism. Perhaps Affirmative Action was needed in the past

Affirmative Action is institutionalized discrimination, at least when used to promote some groups over others. (Though it didn’t start that way; it started as a call to be purely race-blind in hiring.) I wouldn’t call it racism though, because it’s not based on any belief that races have different capability, it is purely intended to correct systemic bias based on the belief that races are equally capable.

it was past due to get rid of it

This might be true, but there are still achievement and pay gaps in the US. There are lots of debates about why, and I don’t want to start one. I’m just curious how else to solve systemic biases if they’re still here. The whole problem with cultural bias is it’s sticky and difficult and people don’t believe they have biases. Today’s politics has done a lot to convince me that we haven’t solved it yet, but at the same time I’ll be the first to point out that we’ve come a long way even in my lifetime. The last little bit might take longer to fix than suffrage did just because of how subtle the issues are. If we take any preferential treatment off the table, preferential treatment that tries to artificially force equal opportunity, the question is what’s the alternative? We might have momentum, and do nothing might work, but what if it doesn’t? Wouldn’t that also be a form of institutionalized discrimination, effectively, like it was before Affirmative Action existed?

Transcribing is transformative.

Not under US Copyright Law, it isn’t. Transcribing is derivative, not transformative.

BTW, if it was transformative, it’d be fair use and legal, and you’d be contradicting yourself.

Today, depending on which model you use. You’re making unstated assumptions. And that’s not a fundamental property of LLMs, it’s happenstance. LLMs are capable of ‘plasticity’, by design.

Scientists who adopt AI gain productivity and visibility: On average, they publish three times as many papers, receive nearly five times as many citations, and become team leaders a year or two earlier than those who do not.

To me this effect doesn’t seem to reflect on AI very much, it seems to reflect on humans. Like maybe this is more evidence of the Babble Hypothesis and the incentives in research than AI, no?

https://en.wikipedia.org/wiki/Babble_hypothesis

always a hobby project, never something serious.

I don’t know what you’re reading, but always and never are strong words. I’ll predict by this time next year you’ll have seen some pretty serious AI uses, and can no longer say always/never. Widespread use of AI coding is brand new, and the models only just barely got good enough to do serious things. It’s way too early to be using words like always and never, but FWIW I’ve already seen some serious uses. There are good reasons personal blog posts rarely talk about ‘serious’ production code; it may be against organizational policy, it may involve code that isn’t’ public, it may reveal proprietary information, and more…

It’s a tool. Good for some things but not for others and generally not to be trusted.

I agree completely you always need to check the work of LLM agents, but it does strike me as a tiny bit funny to anthropomorphize AI by using ‘trust’ while warning against anthropomorphizing the AI by using unchecked output. ;) Generally speaking, “trust” in AI has been going up very quickly as the models & harnesses improve, and as people figure out effective workflows.

I trust my hammer with nails but not screws… does that mean the hammer should generally not be trusted? The problem with AI is we don’t know the difference between nails and screws. (This may be where my analogy breaks down. :P) But I feel like saying don’t trust it isn’t as helpful as saying something like you should expect to spend more time planning and iterating than before, and you should expect tot spend more time reviewing and checking output than before, and learn how to use skills and context and subagents, and learn to use AI on some non-production low-consequence projects first. Saying ‘generally not to be trusted’ implicitly suggests not using AI, and doesn’t leave the reader with how to use AI. The goal is to build trust by building good workflows and by understanding what works well and what doesn’t, right?

the models are quite literally stupid.

You’re arguing via reductionism, and failing to explain the outcomes and emergent properties of the “stupid” system. Humans are made of atoms that are quite literally stupid, so by all means, explain our intelligence and why it’s different than LLMs. (I’m not claiming LLMs are intelligent, BTW, I just don’t think your claim helps nor believe that you can fix it.)

https://en.wikipedia.org/wiki/Reductionism#Definitions